The largest drawbacks in traditional optimization, especially for applications that rely on accessing large datasets, remain significant barriers needing solutions: namely, the sophisticated programmer knowledge that traditional optimization requires, the fragility of tuned solutions in dynamic analyses, and the abundant software development time needed to create an optimized solution. The need remains for improvements in automatically optimizing database analytics operations to improve performance and flexibility while reducing reliance on application programmer skills.